In recent years, one data memory is susceptible to an input request and/or an output request that is issued by plural packets in, for example, a course of execution of communication applications, thereby requiring a division of the data memory into plural use areas for each of the plural packets. In the situation described above, a link list memory is employed to manage the use areas for efficient use of vacancy of the data memory. One of such memory management methods is disclosed in Japanese patent document JP-A-2000-200210.
However, the memory management method by the link list has a problem that data from a specific input request occupies almost all memory areas with the rest of the input requests having no vacant memory areas, because the conventional management method allocates the memory areas in a time-series manner, that is, in a first-come first-served basis.